import re
from textwrap import dedent
from typing import Any, Tuple
from nbconvert.exporters.exporter import ResourcesDict
from nbformat.notebooknode import NotebookNode
from traitlets import Bool, Dict, Unicode
from traitlets.config.loader import Config
from grader_convert import utils
from grader_convert.preprocessors.base import NbGraderPreprocessor
[docs]class ClearSolutions(NbGraderPreprocessor):
code_stub = Dict(
dict(
python="# YOUR CODE HERE\nraise NotImplementedError()",
matlab="% YOUR CODE HERE\nerror('No Answer Given!')",
octave="% YOUR CODE HERE\nerror('No Answer Given!')",
sas="/* YOUR CODE HERE */\n %notImplemented;",
java="// YOUR CODE HERE",
),
help="The code snippet that will replace code solutions",
).tag(config=True)
text_stub = Unicode(
"YOUR ANSWER HERE", help="The text snippet that will replace written solutions"
).tag(config=True)
begin_solution_delimeter = Unicode(
"BEGIN SOLUTION", help="The delimiter marking the beginning of a solution"
).tag(config=True)
end_solution_delimeter = Unicode(
"END SOLUTION", help="The delimiter marking the end of a solution"
).tag(config=True)
enforce_metadata = Bool(
True,
help=dedent(
"""
Whether or not to complain if cells containing solutions regions are
not marked as solution cells. WARNING: this will potentially cause
things to break if you are using the full nbgrader pipeline. ONLY
disable this option if you are only ever planning to use nbgrader
assign.
"""
),
).tag(config=True)
def _load_config(self, cfg: Config, **kwargs: Any) -> None:
if "code_stub" in cfg.ClearSolutions:
if not isinstance(cfg.ClearSolutions.code_stub, dict):
self.log.warning(
"The ClearSolutions.code_stub option must now be given as a "
"dictionary with keys for the language of the notebook. I will "
"automatically convert ClearSolutions.code_stub to a dictionary "
"with a key for 'python', but note that this functionality may "
"be removed in future releases."
)
cfg.ClearSolutions.code_stub = dict(python=cfg.ClearSolutions.code_stub)
if "comment_mark" in cfg.ClearSolutions:
self.log.warning(
"The ClearSolutions.comment_mark config option is deprecated. "
"Please include the comment mark in ClearSolutions.begin_solution_delimeter "
"and ClearSolutions.end_solution_delimeter instead."
)
del cfg.ClearSolutions.comment_mark
super(ClearSolutions, self)._load_config(cfg, **kwargs)
def _replace_solution_region(self, cell: NotebookNode, language: str) -> bool:
"""Find a region in the cell that is delimeted by
`self.begin_solution_delimeter` and `self.end_solution_delimeter` (e.g.
### BEGIN SOLUTION and ### END SOLUTION). Replace that region either
with the code stub or text stub, depending the cell type.
This modifies the cell in place, and then returns True if a
solution region was replaced, and False otherwise.
"""
# pull out the cell input/source
lines = cell.source.split("\n")
if cell.cell_type == "code":
stub_lines = self.code_stub[language].split("\n")
else:
stub_lines = self.text_stub.split("\n")
new_lines = []
in_solution = False
replaced_solution = False
for line in lines:
# begin the solution area
if self.begin_solution_delimeter in line:
# check to make sure this isn't a nested BEGIN
# SOLUTION region
if in_solution:
raise RuntimeError("encountered nested begin solution statements")
in_solution = True
replaced_solution = True
# replace it with the stub, indented as necessary
indent = re.match(r"\s*", line).group(0)
for stub_line in stub_lines:
new_lines.append(indent + stub_line)
# end the solution area
elif self.end_solution_delimeter in line:
in_solution = False
# add lines as long as it's not in the solution area
elif not in_solution:
new_lines.append(line)
# we finished going through all the lines, but didn't find a
# matching END SOLUTION statment
if in_solution:
raise RuntimeError("no end solution statement found")
# replace the cell source
cell.source = "\n".join(new_lines)
return replaced_solution
[docs] def preprocess(
self, nb: NotebookNode, resources: ResourcesDict
) -> Tuple[NotebookNode, ResourcesDict]:
language = nb.metadata.get("kernelspec", {}).get("language", "python")
if language not in self.code_stub:
raise ValueError(
"language '{}' has not been specified in "
"ClearSolutions.code_stub".format(language)
)
resources["language"] = language
nb, resources = super(ClearSolutions, self).preprocess(nb, resources)
if "celltoolbar" in nb.metadata:
del nb.metadata["celltoolbar"]
return nb, resources
[docs] def preprocess_cell(
self, cell: NotebookNode, resources: ResourcesDict, cell_index: int
) -> Tuple[NotebookNode, ResourcesDict]:
# replace solution regions with the relevant stubs
language = resources["language"]
replaced_solution = self._replace_solution_region(cell, language)
# determine whether the cell is a solution/grade cell
is_solution = utils.is_solution(cell)
# check that it is marked as a solution cell if we replaced a solution
# region -- if it's not, then this is a problem, because the cell needs
# to be given an id
if not is_solution and replaced_solution:
if self.enforce_metadata:
raise RuntimeError(
"Solution region detected in a non-solution cell; please make sure "
"all solution regions are within solution cells."
)
# replace solution cells with the code/text stub -- but not if
# we already replaced a solution region, because that means
# there are parts of the cells that should be preserved
if is_solution and not replaced_solution:
if cell.cell_type == "code":
cell.source = self.code_stub[language]
else:
cell.source = self.text_stub
return cell, resources